Identity Document Capture & Quality Guidelines
The NorthRow Mobile Identity Document Verification solution is used by clients looking to undertake AML/KYC checks, Customer Due Diligence, and help streamline the client or employee onboarding journey. It can also be used to validate employees ‘Right to Work’ status.
Our solutions allows the user to simply snap a picture of their driver’s licence, passport or national identity card using their smartphone or tablet. The digital content is then extracted and validated using advanced, automated software to confirm the authenticity of the applicant’s documents.
However, one of the most common issues that we find when processing client document uploads is that of poor image quality, which can result in avoidable delays.
The purpose of this blog is to help clients capture high-quality images of identity documents that can be processed easily, instantly and successfully, without error.
How document processing works?
In addition to our partnerships with a number of leading commercial data providers, NorthRow’s document processing system harnesses the very latest in Artificial Intelligence, Machine Learning and Amazon Web Services (AWS), to ensure we can classify the widest range of documents for our clients.
After 5 years plus of in-depth analysis of processing large volumes of identity documents for our clients, we have turned a wealth of data into a powerful workflow tool, that can make smarter, faster and better informed decisions.
Optical Character Recognition
NorthRow’s Optical Character Recognition (OCR) works by detecting the outline of characters in an image and matching the patterns found against a known set of patterns that represent characters. In order to do this, the image has to be a sharp, high-resolution image where all characters are well defined. When the quality is reduced, this can lead to errors, either individual characters are misread, or the entire image cannot be processed. Our Document Capture Guide explains how to ensure the documents you capture meet these standards.
ID document verification
Once a document has been read, the data is interpreted to determine whether it is a valid identity document. This includes calculation of check digits in Machine Readable Zones (MRZs) and comparison of biometric data to the MRZ. Misreads or missing data can subsequently cause validations error at this point.
General image quality standards
All images must meet these standards. If they do not they may be rejected or the processing may not be successful.
The type of device used to capture an image can vary greatly. It is common that a smartphone or desktop flatbed scanner is used to capture the image, but other devices can also be used. NorthRow recommends that you use a modern smartphone or desktop scanner to capture the image.
Smartphones & Tablets
If you use a smartphone or tablet NorthRow recommends the following specifications:
Minimum 8 MP resolution
Below these capabilities, the image may be hard to capture accurately
Taking a picture
The identity document should be laid on a flat surface and the device held directly above it. The details page should fill the screen, and the image should be in focus. There should not be any glare or other images in the shot.
If a scanner is used NorthRow’s recommends the following specifications:
Minimum 300 DPI scanning resolution
Save as an image, for example, JPEG or PNG not PDF. PDF file formats can cause processing issues, described later
The document type is set to photo, if available
Scanning an image
Lay the document flat in the top left corner of the scanner. Do not place a sheet of paper behind the document as this may result in the scanner creating an image with a large white border. Scan the document and save the image.
Types of Document
Most documents have the only page that needs to be captured, for example, the detail page of a passport. However, some documents are double sided, for example, a Biometric Residence Permit. If you need to capture a double sided each side must be a separate image. These can then be submitted to NorthRow as a set.
If you need to capture a passport and supporting visa each document should be captured and submitted as individual documents in a check.
We do not recommend editing images before submitting them for processing. Applications such as PhotoShop will leave additional information in the file that indicates it has been edited and this may cause the image to be rejected as suspicious.
NorthRow recommends that you submit document images in one of the following formats
We also accept PDF documents. However, PDF can result in lower success rates as the image may be altered within the document. This is due to the nature of the file format which can result in images being broken up into many smaller parts. This makes it impossible to process automatically. Where possible NorthRow will convert PDFs into images before processing.
Images can be submitted to NorthRow through the GUI or API. Please contact firstname.lastname@example.org for an overview of submitting documents through the GUI, or for further information on our API